Sepsis kills more than 215,000 Americans each year. It is estimated that 750,000 Americans are infected with severe sepsis and 29% of them die from it each year. Sepsis deaths make 9% of all death cases in the U.S. Sepsis kills as many Americans as myocardial infections, even more than traffic accidents.
Two to three million Americans acquire a hospital infection each year and 10% of these infections progress to sepsis. More than 90,000 of these patients die from sepsis infected in hospitals.
Severe sepsis and septic shock (severe sepsis combined with low blood pressure) took up to 135,000 lives each year in the intensive care units (ICU) in the European Union according to the OECD Health Report of 2000. In Britain, 5,000 out of 100,000 patients who acquired a hospital infection die from sepsis every year in acute care hospitals belonging to the NHS organisation.
The death toll has increased year after year due to the fact that the number of patients predisposed to sepsis, such as the elderly, premature neonates, and cancer patients, has increased, not least because many serious illnesses are more treatable than before. Also the use of invasive medical devices and aggressive procedures has increased.
Gram-negative bacteria cause more than 40% of all septicemic infections and many of the Gram-negative bacteria are extremely multiresistant. Gram-negative bacteria provide a harder challenge in therapy than Gram-positives, as they possess a unique structure, the outer membrane, as their outermost structure. Lipopolysaccharide molecules located on the outer membrane inhibit the diffusion of many antibacterial agents deeper into the cell, where their ultimate targets are located. More than 95% of the novel antibacterial agents isolated from nature or chemically synthesized in 1972-1991 lacked activity against Gram-negatives (Vaara 1993).
Polymyxins are a group of closely related antibiotic substances produced by strains of Paenibacillus polymyxa and related organisms. These cationic drugs are relatively simple peptides with molecular weights of about 1000. Polymyxins, such as polymyxin B, are decapeptide antibiotics, i.e., they are made of ten (10) aminoacyl residues. They are bactericidal and especially effective against Gram-negative bacteria such as Escherichia coli and other species of Enterobacteriaceae, Pseudomonas, Acinetobacter baumannii, and others. However, polymyxins have severe adverse effects, including nephrotoxicity and neurotoxicity. These drugs thus have limited use as therapeutic agents because of high systemic toxicity.
Polymyxins have been used in the therapy of serious infections caused by those bacteria, but because of the toxicity, their use was largely abandoned in the 70's when newer, better tolerated antibiotics were developed. The recent emergence of multiresistant strains of Gram-negative bacteria has necessitated the therapeutic use of polymyxins as the last resort, in spite of their toxicity, and as many of the less toxic antibiotics have already lost their effectiveness against particular strains of the said bacteria, the use of polymyxins has again increased.
Accordingly, polymyxins have now been recalled to the therapeutic arsenal, although, due to their toxicity, on a very limited scale. Their systemic (i.e., non-topical) use is, however, largely restricted to the therapy of life-threatening infections caused by multiply resistant strains of Ps. aeruginosa and A. baumannii as well as by carbapenem-resistant enteric bacteria.
Polymyxins consist of a cyclic heptapeptide part and a linear part consisting of a tripeptide portion and a hydrophobic fatty acid tail linked to the α-amino group of the N-terminal amino acid residue of the tripeptide and may be represented by the general formula:

wherein R1-R3 represent the tripeptide side chain portion; R4-R10 the heptapeptide ring portion and R(FA) represents the hydrophobic fatty acid tail linked to the α-amino group of the N-terminal amino acid residue of the tripeptide.
The polymyxin group includes the following polymyxins: A1, A2, B1-B6, IL-polymyxin B1, C, D1, D2, E1, E2, F, K1, K2, M, P1, P2, S, and T (Storm et al. 1977; Srinivasa and Ramachandran 1979). All polymyxins are polycationic and possess five (5) positive charges, with the exception of polymyxin D, F, and S which possess four (4) positive charges. It should be noted that modified polymyxins that lack the fatty acid part R(FA) but carry R1-R10 have one additional positive charge when compared to the natural polymyxins they derived from, due to the free α-amino group in the N-terminus of the derivative. Accordingly, for example, such a derivative of polymyxin B or polymyxin E carries six (6) positive charges in total.
The clinically used polymyxin B and polymyxin E differ from each other only in the residue R6, which is D-phenylalanyl residue in polymyxin B and D-leucyl residue in polymyxin E.
Also circulin A and B are classified as polymyxins (Storm et al. 1977). They differ from other polymyxins only in carrying isoleucyl residue in the position R7 whereas other polymyxins have either threonyl or leucyl residue in the said position. For an overview of the structures of some polymyxins, see Table 1.
TABLE 1The structure of selected polymyxins and octapeptin as well as selected derivatives thereofCompoundR(FA)R1R2R3R4R5R6R7R8R9R10Polymyxin BMO(H)A-Dab-Thr-Dab-*Dab-Dab-D-PheLeu-DabDab*ThrColistin (polymyxin E)MO(H)A-Dab-Thr-Dab-*Dab-Dab-D-LeuLeu-DabDab*ThrColistin sulphomethateMO(H)A-sm-DabThr-sm-Dab*Dab-Sm-DabD-LeuLeu-sm--Dabsm--Dab*ThrPolymyxin AMO(H)A-Dab-Thr-D Dab-*Dab-Dab-D-LeuThr-DabDab*ThrPolymyxin MMOADab-Thr-Dab-*Dab-Dab-D-LeuThr-DabDab*ThrPolymyxin DMO(H)A-Dab-Thr-D Ser-*Dab-Dab-D-LeuThr-DabDab*ThrCirculin AMOADab-Thr-Dab-*Dab-Dab-D-LeuIle-DabDab*ThrOctapeptin AOHMDA—Dab-*Dab-Dab-D-LeuLeu-DabDab*LeuDeacylcolistin (DAC)Dab-Thr-Dab-*Dab-Dab-D-LeuLeu-DabDab*ThrPolymyxin E nonapeptide (PMEN)Thr-Dab-*Dab-Dab-D-LeuLeu-DabDab*ThrDeacylpolymyxin B (DAPB)Dab-Thr-Dab-*Dab-Dab-D-PheLeu-DabDab*ThrPolymyxin B nonapeptide (PMBN)Thr-Dab-*Dab-Dab-D-PheLeu-DabDab*ThrPolymyxin B octapeptide (PMBO)Dab-*Dab-Dab-D-PheLeu-DabDab*ThrPolymyxin B heptapeptide (PMHP)*Dab-Dab-D-PheLeu-DabDab*Thr
Polymyxin B is represented by the following formula:

Commercially available polymyxin B is a mixture, where R-FA is predominantly 6-methyloctanoyl (6-MOA, in polymyxin B1) but may also be a related fatty acyl such as 6-methylheptanoyl (6-MHA, in polymyxin B2), octanoyl (in polymyxin B3), or heptanoyl (polymyxin B4) (Sakura et al. 2004). All these variants are equally potent against Gram-negatives such as E. coli (Sakura et al. 2004). Quite analogously, in polymyxin E1 (colistin A) and in circulin A the R-FA is 6-MOA and in polymyxin E2 (colistin B) and in circulin B the R-FA is 6-MHA. Numerous researchers have attached various hydrophobic moieties including various fatty acyl residues to the N-terminus of polymyxin derivatives and analogues and have shown that the resulting derivatives have potent antibacterial activity (Chihara et al. 1973, Sakura et al. 2004 and in US patent publication 2006004185. Even the derivative that carries the bulky hydrophobic 9-fluorenylmethoxycarbonyl residue as the R-FA is almost as potent as polymyxin B in inhibiting the growth of E. coli and other Gram-negative bacteria (Tsubery et al. 2001).
For biological activity the heptapeptide ring structure is essential (Storm et al. 1997). A derivative with an octapeptide ring is significantly less active as an antibiotic.
Multiple modifications of polymyxins and multiple polymyxin-like synthetic molecules have been made, and with certain limits they have preserved their biological activity. The modifications comprise but are not limited to those in the side chain, as well as molecules in which an inherent hydrophobic amino acid residue (such as DPhe or Leu) has been replaced with another hydrophobic amino acid residue or in which the cationic Dab has been replaced with another cationic amino acyl residue, such as Lys, Arg, or ornithine residue (Storm et al. 1997, Tsubery et al. 2000a, Tsubery et al. 2002, US patent publication 2004082505, Sakura et al. 2004, US patent publication 2006004185).
Other modifications that result in microbiologically at least partially active compounds comprise but are not limited to alkanoyl esters where the OH-groups of the threonyl residues form esters with alkanoyls such as propionyl and butyryl (U.S. Pat. No. 3,450,687).
Octapeptins are otherwise identical to polymyxins but have a covalent bond instead of the residues R1-R2 (Table 1). In this invention, the R positions are numbered according to those in the natural polymyxins and thus the only amino acyl residue in the side chain of octapeptins is defined as R3. Accordingly, octapeptins are octapeptides whereas all natural polymyxins are decapeptides, and they possess only four (4) positive charges. The R-FA residues among various octapeptins (A1, A2, A3, B1, B2, B3, C1) include the following: 3-OH-8-methyldecanoic acid, 3-OH-8-methylnonanoic acid, and β-OH-6-methyloctanoic acid. Derivatives that possess a fatty acyl residue with 6 to 18 carbon atoms have a potent antibacterial activity against E. coli (Storm et al. 1977).
The first target of polymyxins in Gram-negative bacteria is their outer membrane (OM) that is an effective permeability barrier against many noxious agents including large (Mw more than 700 d) antibiotics as well as hydrophobic antibiotics. By binding to the lipopolysaccharide (LPS) molecules exposed on the outer surface of the OM, polymyxins damage the structure and function of the OM and, as a result, permeabilize (i.e., make permeable) the OM to polymyxin itself, as well as to many other noxious agents (Nikaido and Vaara 1985, Vaara 1992, Nikaido 2003). The final and lethal target (the bactericidal target) of polymyxins is believed to be the cytoplasmic membrane (the inner membrane) of bacteria.
Numerous efforts have been made to reduce the toxicity of polymyxins. The treatment of polymyxin E (colistin) with formaldehyde and sodium bisulfite yields colistin sulphomethate, in which the free amino groups of the five diaminobutyric acid residues have partially been substituted by sulphomethyl groups (Table 1). The preparations consist of undefined mixtures of the mono-, di-, tri-, tetra-, and penta-substituted compounds. The sulphomethylated preparations, when freshly dissolved in water, initially lack both the antibacterial activity and toxicity of the parent molecule, but when the compounds start decomposing in the solution, in the blood or in the tissues to yield less substituted derivatives and free colistin, both the antibacterial activity and the toxicity are partially brought back. Furthermore, the degree of initial sulphomethylation apparently varies between the commercially available pharmaceutical preparations. Many other ways to block all the free amino groups have been published. Examples comprise but are not limited to the formation of unstable Shiff bases with amino acids (Storm et al. 1977).
Polymyxin E nonapeptide (PMEN, colistin nonapeptide, Table 1), obtained by treating polymyxin E enzymatically and lacking the R-FA and R1, was shown in 1973 to be less toxic than the parent compound in acute toxicity assay (immediate death presumably due to direct neuromuscular blockade) in mice (Chihara et al. 1973). However, it also lacked the antibacterial activity, as measured as its ability to inhibit bacterial growth (Chirara et al. 1973). The role of the linear part may contribute to the antibacterial activity of the polymyxins.
Vaara and Vaara, on the other hand, showed that polymyxin B nonapeptide (PMBN, Table 1) retains the ability to permeabilize the OM of Gram-negative bacteria (Vaara and Vaara 1983a,b,c; U.S. Pat. No. 4,510,132; Vaara 1992). Accordingly, even though it lacks the direct antibacterial activity (i.e., the ability to inhibit bacterial growth), it is able to sensitize (i.e., make sensitive or, as also termed, make susceptible) the bacteria to many antibacterial agents such as hydrophobic antibiotics as well as large antibiotics and some other noxious agents.
PMBN also sensitizes bacteria to the bactericidal activity of the human complement system, present in fresh human serum as a first-line defence system against invaders (Vaara and Vaara 1983a, Vaara et al. 1984, Vaara 1992). Furthermore, it sensitizes the bacteria to the joint bactericidal activity of serum complement and human polymorphonuclear white cells (Rose et al. 1999).
PMBN resembles PMEN in being less toxic in the acute toxicity assay in mice than unmodified polymyxins. In further toxicological assays, several criteria proved PBMN to be less toxic than its parent compound, but this polymyxin derivative was still judged to be too nephrotoxic for clinical use (Vaara 1992).
PMBN carries five (5) positive charges. Subsequent studies revealed, quite expectedly, that PMEN, also carrying five (5) positive charges as well as deacylpolymyxin B and deacylpolymyxin E, both carrying six (6) positive charges are potent agents to sensitize bacteria to other antibiotics (Vilijanen et al. 1991, Vaara 1992). In addition, it has been shown that a structurally further reduced derivative polymyxin B octapeptide (PMBO) retains a very effective permeabilizing activity while polymyxin B heptapeptide (PMBH) is less active (Kimura et al. 1992). PMBN, PMEN and PMBO have five (5) positive charges while PMBH has only four (4) positive charges. This difference may explain the weaker activity of PMBH.
The group of Ofek, Tsubery and Friedkin recently described polymyxin-like peptides that were linked to chemotactic peptides, such as fMLF, that attract polymorphonuclear leucocytes (US patent publication 2004082505, Tsubery et al. 2005). They described peptides fMLF-PMBN, MLF-PMBN, fMLF-PMEN, fMLF-PMBO and MLF-PMBO, all carrying four (4) positive charges, that sensitize Gram-negative bacteria to antibiotics, even though no comparative studies with increasing concentrations of the compounds were published (Tsubery et al. 2005).
In order to study the structures and functional properties of polymyxins, a few works have disclosed, among other compounds, polymyxin derivatives having less than four (4) positive charges.
Teuber (1970) has described the treatment of polymyxin B with acetic anhydride that yields a preparation containing polymyxin B as well as its mono-, di-, tri-, tetra-, and penta-N-acetylated forms. Teuber also separated each group and nonquantitatively reported using an agar diffusion assay that penta-acetylated and tetra-acetylated forms lacked the ability to halt the growth of Salmonella typhimurium, whereas di- and monoacetylated forms did have such ability. Triacetylated form had some ability.
Srinivasa and Ramachandran (1978) isolated partially formylated polymyxin B derivatives and showed that a diformyl derivative as well as a triformyl derivative inhibited the growth of Pseudomonas aeruginosa. They did not disclose the compounds' ability to sensitize bacteria to antibiotics. Furthermore, in 1980 they showed that the free amino groups of triformylpolymyxin B in residues R1 and R3, as well as the free amino groups of diformylpolymyxin B in residues R1, R3, and R5 are essential while the free amino groups in R8 and R9 are not essential for the growth inhibition (Srinivasa and Ramachandran, 1980a).
A shortened polymyxin B derivative octanoyl polymyxin B heptapeptide has been disclosed by Sakura et al. (2004). The attachment of the octanoyl residue to the N-terminus of the residue R4 of the polymyxin B heptapeptide results in a compound having only three (3) positive charges. Sakura et al. found that octanoyl polymyxin B heptapeptide inhibits the growth of bacteria only at a very high concentration (128 μg/ml), whereas the other derivatives such as octanoyl polymyxin B octapeptide and octanoyl polymyxin B nonapeptide, both having four charges (4) were very potent agents to inhibit bacterial growth.
US patent publication 2006004185 recently disclosed certain polymyxin derivatives and intermediates that can be used to synthesize new peptide antibiotics. The antibacterial compounds described possessed four (4) or five (5) positive charges.
Furthermore, closely related polymyxin B and polymyxin B1 compounds have also been disclosed by Okimura et al. (2007) and de Visser et al. (2003). Okimura et al. have studied the chemical conversion of natural polymyxin B and their N-terminal derivatives and de Visser et al. have studied solid-phase synthesis of polymyxin B1 and analogues via a safety-catch approach. The antibacterial compounds described in these works possessed four (4) or five (5) positive charges.
There is still an urgent need for effective treatments for bacterial infections, in particular for the infections caused by multiresistant Gram-negative bacteria.